Remodeling RPGs for the New Millennium

Despite working in a truly remarkable medium, one that provides powerful tools for the simulation of fantastic worlds and myriad ways to immerse players in them, many RPG designers these days seem content to recreate the glories of earlier computer games. Worse, many designers seem content to recreate experiences that they (and we, as players) first enjoyed in other media.

As I stated in my recent Soapbox column ("It's ROLE-playing, Stupid!", Game Developer magazine, September 1998), we RPG designers are setting our sights too low. Look at the best RPGs of the last several years. As great as Diablo, Fallout, Daggerfall, and Might & Magic VI are, they really aren't anything that we couldn't have designed ten years ago. Do they represent significant advances over Wasteland or Ultima IV or the Underworld games? And were these older games striving for much more than a recreation of the tabletop role-playing experiences of their creators? It's as if we can't see beyond our early Dungeons & Dragons game experiences. It's time to move beyond simply borrowing game concepts and establish computer RPGs as an independent medium.

Here are some things I won't be doing in this article:

So that's what I won't be doing. What am I going to do?

First and foremost, I want to prod you into thinking a bit more about your design choices. I don't want you to feel comfortable doing what you've done before simply because that's the way you've always done it (or the way Dave Arneson and Gary Gygax did it back in the Bone Ages). It's time to break the cycle, to acknowledge that we're in a rut. Time to think through our design decisions, to know why we're making the choices we make. Time to think through the ramifications of those choices.

My intent with this article is to lay out the abundant variety of choices available to would-be and practicing RPG designers. Only by analyzing the tools that we all use in the creation of our games, discussing the ways in which these tools have been and can be used, and identifying the ramifications of those uses, can we take this genre forward.

Fundamentally, this article is an attempt to identify what it's going to take to be a player in the RPG category in the years to come. I hope to identify the critical elements a game must have to compete and dominate. However, before looking at what it will take to best the RPG competition in 1999 and beyond, it seems reasonable to look (briefly) at where roleplaying is today and how we got here.

So, how did we get to where we are today in RPG development? What are the catalysts for rising public interest and publisher confidence in a category once considered too "nichy" to bother supporting? It all started with the release a few years ago of some surprisingly successful PC RPGs, notably Diablo and Daggerfall. After years of climbing costs, lengthening development cycles and steady, unspectacular sales, these two games broke out of the niche. At the same time, new game consoles appeared, and RPGs have always been a factor in the console market. Throw in the still on-going convergence of the PC and console markets, and roleplaying looks a lot more attractive -- and potentially profitable -- than it has in years.

That's a brief look at why the future looks rosy. But that begs the question, if things are that promising, why does the current crop of RPGs look so old fashioned? To answer that question, we have to look at computer roleplaying's beginnings -- at our earliest inspirations as designers.

The roots of computer roleplaying stretch way back. Some see the genre's beginnings in the make-believe games played by children since time immemorial (with cowboys and indians being the canonical example in the United States). Others point to the re-issue of Tolkein's Lord of the Rings in the '60s as our field's genesis. These games and novels surely influenced us all. But the fact is that none of you reading this article would have jobs were it not for Dave Arneson and Gary Gygax, creators of the game "Dungeons & Dragons". A history of D&D would be a distraction here, and others have covered that territory far better than I could. But our debt to the creators of D&D is worth noting here if only to force a recognition of how little we've moved beyond the realm of 20-sided dice, the concept of character class and those incredible core attributes we can all rattle off like a mantra. (Come on, recite them with me: STR, DEX, CON, INT, WIS, CHA. And if you try to tell me you don't know what those abbreviations stand for, you're either a liar or you should drop this magazine immediately and get back to the latest exciting issue SQL Database Programmers Journal or something.) If you've ever worked on a computer RPG, you've ripped off Arneson and Gygax at some point. Admit it, but don't feel bad about it. As pioneers in the birth and development of a new medium, your borrowings put you in good company.

How do we identify a computer RPG? For the purposes of this article, a computer RPG is a game in which character development and character interaction take precedence over other factors and where each player's experience of the story is determined by individual choice rather than designer fiat. Though broad, this definition clearly eliminates real-time and turn-based puzzle and strategy games (lack of character development and interaction), as well as shooters and platform action games (lack of individual choice). Of greatest importance, this definition eliminates adventure games, which share with the RPG an emphasis on story and character. What adventure games lack - and this is a critical point - is the capability for players to grow and develop their characters, and to affect, if not the outcome of the story, than the way in which the story unfolds. Without both character development and genuine choice placed within a player's control, a game cannot be called a role-playing game, as I choose to define the genre.

Someday, we will concentrate on those aspects of computer RPGs that set them apart and we will leave our paper gaming roots behind. To do so, we have to be more daring in our designs - or, in the terminology of this article, in our selection of design tools. And to select the right tools, we need a better understanding of these tools and how they define the genre.

The term "tool" seems, at first, odd to use in the context of game design. When I use this term, I'm talking about the conventional elements that are sure to appear in any work that defines itself, or is defined, as a computer RPG. Any designer contemplating an RPG must take a stance with regard to all of these tools, even if that stance is to deemphasize one or more of them. What follows are the defining characteristics that must be present in any RPG.

Story

Story RPGs are story driven. There's a reason for talking to or killing people and monsters, a reason to build or destroy things. Unfortunately, though it goes without saying that RPGs must tell a tale, it's unclear whether the quality of that tale has much, if anything, to do with a game's success. One would be hard-pressed to describe the Avatar vs. Guardian (a.k.a. "kill the evil foozle") stories of recent Ultima games as on a par with what we demand from books and movies. Diablo's plot hardly qualifies as compelling. Underworld's story of a hero locked in a dungeon until he can rescue a kidnapped princess hardly qualifies as narrative genius.

Currently, the kinds of stories we can tell seem to be limited by the expressive capabilities of our medium - it's tough to tell a great story when you -can't recreate a young lover's shy smile or allow players to tell a joke rather than bludgeon somebody. Right now, what we do most easily and best is direct, one-on-one conflict (typically combat interactions), a fact that limits our narrative range just a tad. This is not to say we shouldn't strive for greatness in our stories, but we must find greatness in the strength of universal themes and in the ways in which we tell our necessarily simple stories.

Players of RPGs must have some degree of freedom in how they follow the threads of the plot and, in some cases, how the plot resolves itself. They can often pick the order in which they accept quests or even which quests they take and which ones they ignore. Further, how they conduct themselves during a quest, and how they interact with other characters, can alter the course of the story and its outcome.

The first and, arguably, most critical decision RPG developers must make with regard to story is whether to use a branching structure or to tell a story in a more conventional, linear fashion. The temptation is strong simply to say, "We're making a computer game. Computers allow branching in a way and at a level no other medium allows. Of course, we should use a branching structure." This argument, one I've made myself, goes back to the moral imperative to maximize the unique capabilities of the medium and to turn away from the techniques more appropriate to other media. It's perfectly understandable that computer RPG developers would want a branching structure if for no other reason than to differentiate games from books and movies. But let's think through the implications of that decision.

Often, making one choice - picking one branch over another - means that a player can't go back to the branch not chosen. If I may be prescriptive for a moment, if picking a branch does not limit players' later options in some way, the branch is unnecessary and a waste of valuable development time. The illusion of player freedom isn't worth the development price.

However, assuming branching offers real choices (meaning, choices that limit player options even as the player moves forward through the plot), the approach can be worth the cost. Done well, branching can provide a powerful illusion of freedom for players. But, that's all it can provide - an illusion. The reality is that, if we don't put something in the game, on the screen, in the mouths of nonplayer characters (NPCs), it doesn't happen - and no amount of branching can allow players to do things we don't allow them to do. What this means is that the choices available to players solely as a result of branching are false, because eventually players are forced back onto one of the paths that we've created for them.

The first factor to consider when assessing whether branching is appropriate and/or necessary for your project is whether it's worth sinking valuable development resources into the creation of content that many, if not most, players will never see. And bear in mind that you're going to be spending time and money to ensure that the game makes sense regardless of the order in which each player sees each portion of the story. That's a lot of extra flags to set and check and a lot of extra art to create on the off-chance that players will stray from the logical path.

But what about replayability? Doesn't branching encourage players to keep playing a game? My first response would be, "Nah. By the time they finish your 100-hour epic, they're probably looking for The Next New Thing." Only the most zealous players replay games at all, and they're sure to see that a big percentage of their adventure differs not at all on subsequent playthroughs no matter which plot branches they follow.

None of this is to say that branching -isn't worth all that extra effort. Though not vital to success (aesthetic or commercial) it's important that players talk about their experiences playing your RPG and, when they do, it's powerful when their descriptions differ, seemingly based on individual choices. As in all development-oriented decisions, it's important to weigh that power against the cost of achieving it.

It's also important to realize that once you do spend your development dollars on giving the player power over the way in which your story unfolds, that should become the emphasis of your game. You should try to give your players a big, contiguous world to explore and you should let them explore it freely and in wany way they want - even at the expense of character development.

The alternative to branching is to tell a more traditional linear story. But telling a story in the way that stories have always been told isn't the answer. So what are the advantages of telling a linear story and how is this best achieved? Let's start with the biggest and most obvious advantage of the linear narrative, the story itself.

Clearly, you can tell a better story if you don't have to worry about and/or deal with all the ways in which players can screw up your carefully crafted epic narrative. It's generally accepted that a linear story in a game almost inevitably means a more powerful story. Given the cost of achieving the illusory freedom offered by branching storylines, the linear story seems to be a pretty good deal. In addition, depending on how you implement your linear story, you may find it possible to give players some genuine freedom to personalize their experience rather than the illusion of freedom offered by branching narratives and huge worlds to explore.

What I'm getting at is that a linear story must have two characteristics. As the creator of a linear RPG, you must offer the player flexibility within episodes or narrative segments or on a single map or within a single mission. Combine this flexibility with a focus on something other than narrative (such as character development) as the driving principle behind your game, and players won't notice that they're on rails, narratively speaking.

Final Fantasy VII does a wonderful job of allowing you to explore each of its locations with some degree of freedom. Players rarely feel constrained or stuck to a path, even though they are. The reason lies in the game's emphasis on character development. The designers recognized that freedom of movement would eventually interfere with the advancing plot, so they emphasized systems that allow players to create unique alter egos who respond to scripted events in ways that are often within the player's control. This feature allowed them to tell a better story with more interesting characters than would be present in a nonlinear game. I'm not saying that Final Fantasy is necessarily a better game than Daggerfall (a nonlinear game if there ever was one) - just that the designers clearly thought through the implications of the critical design decision to tell a linear story.

Here, as in most design decisions, there's no right or wrong answer. Linear narratives, expertly implemented, are no better or worse than branching narratives implemented equally well. However, it's worth pointing out that perceived freedom is more important than actual freedom. If the players thinks they're in control, it's as good as if they are.

Character Differentiation & Development

RPGs are character-driven. Unlike any other game genre, they rely on differentiated player characters. As such, unique, personal character growth is vital. Players must feel that they control the destiny of their alter egos and that their choices throughout the game result in increasing stature and a growing ability to impact the game world and its denizens.

Every design decision you make when crafting an RPG should first be filtered through the following simple screens:

If we as game designers allow each player's character to be unique, and thus differentiate each player's experience of the game, we have been successful. To illustrate how important the need to play a role is in role-playing games, and how controversial the subject can be, let me describe some personal experiences. In recent months, I found myself embroiled in a controversy that I never could have imagined. The issue involved the nature of role-playing and character identification.

It occurred at Ion Storm, where I'm currently working on the game Deus Ex. My development team, which is fairly united on role-playing design issues, suddenly found itself on the brink of civil war over whether players should be allowed to name their characters. My original plan had been to give the character a name and a backstory to go along with it. That would allow us to give the character significant relationships and, perhaps most important, a voice.

Half of the team felt that the predetermined name and identity offered too many dramatic advantages to pass up, particularly nowadays when full speech is expected and voice synthesis technology is still in its infancy. The other half of the team was appalled. "If you can't name your character," said one developer, "you're not making an RPG at all. You're making an adventure game." Several people commented that they find it annoying when they are forced to do or say things because the designer thinks their character would do or say that thing. To cut short this debate, I came up with a solution that, I believe, satisfied both camps. (You can tell me how successful my solution was when the game ships!) In any event, this argument about character names shows just how critical player identification with his or her character can be to the success of an RPG.

Statistics, Skills, and/or Trackable Abilities

As tools, names are useful but not critical. In contrast, there are two core character identification and development tools: statistically driven and experientially driven story and world building. Regardless of whether you allow players to develop their characters through statistics or direct experience (or some combination of the two), you have to take a stance on the subject before you put the last period on your design document.

In games such as Fallout, Diablo, Daggerfall, Final Fantasy VII, and Might & Magic VI, numbers typically define your character's fundamental attributes (such as strength, dexterity, intelligence, and luck) and/or your character's level of accomplishment in a set of skills (such as lock picking, marksmanship, and first aid). When players run into a game problem or obstacle, they use one or more of these attributes or skills, resulting in behind-the-scenes die rolls that determine success or failure in overcoming the problem.

Development (increases in individual statistics) often comes through the expenditure of abstract skill points given by the designer for solving individual problems or for solving enough such problems to go up an arbitrary level (the rewards model). In other games, development comes through the actual use of specific capabilities in game situations (the practice model). However implemented, statistics are terrific tools for setting one character apart from another - there's a reason they've been a staple of role-playing since the birth of D&D.

Why are numeric systems terrific tools? For one thing, they're instantly parsable by normal human beings. Any player can tell immediately that first level isn't as good as second level, and that a strength score of 65 is better than a score of 37. In addition, die-rolls introduce tension, suspense, and variety into computer RPGs.

Statistical systems do have associated costs, though. The one that I find most damning, if only because a thoughtful designer can so easily avoid the problem, is that typically, by game's end, characters tend to end up looking more alike than different. But the trick to avoiding this is simply to impose limits on the number of skills players can select and/or to limit the number of reward points we hand out.

But using statistics poses other problems, too. As easy as it is to say two characters are different - and as easy as it is to indicate these differences on a character description screen - it is extremely difficult to communicate to players the reasons why they succeeded or failed at a given task. Can players ever really know why they succeeded or failed when behind the scenes die rolls determine success of failure? Can we make players feel their contributions to character accomplishment are significant? If you choose a statistical approach, you need to provide obvious and immediate feedback when a statistic affects problem resolution.

Games such as Wasteland and Diablo, though separated by many years, are extremely good examples of games that use statistics and skills effectively. In these games, statistics, in addition to being rewards, allow players to refine their characters with a great deal of control and precision and to individualize their play experience in ways the games' limited storylines don't offer. Fallout is a more recent example of effective use of statistics to differentiate characters from one another (Figure 2). But Fallout takes the idea of tailoring experience through statistics even further than Wasteland and Diablo - a player who puts his or her points into stealth and communication skills, for example, is likely to solve game problems very differently than one who puts those points into weapons skills.

Looking at most games that use the statistical character definition approach, you quickly get the impression that designers like to track numerous character statistics, and like to track them to a fine degree. Logically, this seems to be the best way to differentiate one character from another. However, if you're going to track statistics based on skill values, adjustments to these values have to be meaningful enough so that changing them makes obvious changes in the game play. Hand out too many statistical improvement rewards and characters start to look more alike than different. Create enough statistics and skills and players quickly figure out which ones matter and which don't, causing characters to look more similar. My advice is to be appropriately and thoughtfully stingy with rewards and with the number and types of statistics you provide.

But are all of these statistics really necessary? Of course not. There's another way. In recent years, a small, very vocal and extremely persuasive minority of the design community has begun to argue in favor of statistics-free RPGs or, as some call it, the immersive experience. They feel that hidden die rolls and finely tracked statistics are unnecessary hold-overs from paper gaming. These designers pose a number of interesting questions. Why use a crutch from another medium, one with limited simulation capabilities, in computer gaming, which has far more powerful simulation tools available? Why not let player choices determine character differences? Does anyone think the difference between a 17 and an 18 in strength or between an 89 and a 90 in lock picking should have an impact on game play?

So what do these statistics-foes offer as an alternative in terms of character differentiation and the player's ability to impact a story? The two most important alternative tools available are Inventory and Skills/Special Abilities.

Most, if not all, RPGs support the accumulation and use of items by players. In most, you can pick up anything that isn't specifically nailed down and use it later, possibly even in ways the designers never imagined. In extreme cases, you end up with characters hauling around useless candy wrappers and soda cans. In the end, problem resolution is what RPGs are all about. The more tools you give the player (useless items notwithstanding), the more solutions are likely to suggest themselves, as long as your simulation is robust enough - or your designers clever enough - to support them.

In addition to their use in problem solving, objects and weapons can be powerful tools for character differentiation. If you load up your inventory with weapons and I load mine up with keys, lockpicks, and inviso-suits, our characters look, feel, and, of necessity, behave differently from one another.

The key to making inventory a character differentiation tool is to limit, in some way, the number of items that a character can carry. In a statistics-based system, this can be accomplished by giving items weight and then tying inventory capacity to strength - how much weight a character can carry therefore becomes the limiting factor. In a statistics-free system, the same goal can be accomplished by giving each item a size and then limiting the number of things a character can lug around. Clearly, combinations of these ideas work, too - witness Diablo (Figure 3) - and there are undoubtedly several other viable schemes.

For inventory to work as a character and experience differentiation tool, we must find ways to force players to make choices. Which implies limiting characters' capabilities. We as designers must be disciplined enough to parcel out items of increasing power - things that make characters more effective - in a careful, well thought-out manner. If you dumped an infinite variety of weapons in front of a character, most players would grab the most powerful one, making the inventory limitation moot. Item and weapon differentiation must be thought of in terms of economy.

In a real-world economy, more -isn't always better. The same is true in gaming. Just because you can offer players 4,000 weapons doesn't mean that you should. The choice should have meaning. Ask yourself whether you can really differentiate two weapons; if the two items offer no legitimate, significant, and obvious game play difference in your game world, why bother? When pondering inventory issues, think of yourself as the game design equivalent of Alan Greenspan of the Federal Reserve Board - you have to open and close the object floodgates to match players' capabilities to the tasks at hand. Release too much "item power" and tasks become too easy. Interest in the game wanes. Release too little "item power" and tasks become too hard. Frustration sets in. Either way, players stop playing.

Not all systems of skills and abilities depend on statistical resolution; there are plenty of other ways to structure a system. In Deus Ex, we use a binary action resolution scheme in which your skill level, tracked in a gross fashion (rather than a granular one) is compared to the difficulty of the task at hand. If your skill level is higher than the difficulty factor - which, in most cases, you'll know before you attempt a task - then you succeed. If your skill level is lower than the task's difficulty factor, you have to find another solution.

In a deeply simulated world (or even a modestly simulated one, such as Deus Ex) each game problem should be solvable in a variety of ways. In such a simulation, a skill system like our binary action resolution model makes perfect sense. If you're not a good enough lock picker to open a vault, maybe you're a master with explosives, or maybe you can charm a bank clerk into opening it for you. If you're still thinking in terms of puzzles rather than obstacles, and if your world is two-dimensional rather than deeply simulated, stick with statistics.

So where does tension come into the picture? To offer levels of suspense like statistics-based systems, a statistics-free game must emphasize consequence and reward. Here's one way it can work. First, players may know in advance the outcome of a specific action (whether they do or not is largely a result of how much attention they've been paying - the designers aren't trying to hide anything). Second, players should have a fair degree of certainty about the reward for acting in a particular manner (for example, a player should know that if he or she gets through the vault door, he or she will get a million dollars). Most critically, players must not be able to do more than make an educated guess at the consequences of acting in a specific manner (for example, picking the lock on the vault door might or might not set off an alarm, blowing the door off it's hinges might attract the attention of the night watchmen or destroy the money in the vault, and charming a bank teller might allow the teller to identify you when the police show up). All actions - all choices - must reward players and, equally important, all must have consequences. There can be no right and wrong, no better and worse.

Whether statistics or experience is used to differentiate characters, some tangible measurement of character prowess and progress is necessary if a game is to be considered an RPG. In RPGs, arbitrary limitations are often placed on what your character can and cannot do. The idea of defining your character's abilities statistically is just one such arbitrary limitation.

Another is the notion of the character class. In the past, distinguishing characters using character classes has prevented mages from wearing armor, impaired warriors' use of magic, restricted clerics' use of offensive magic, and let thieves move quietly and do double damage from behind. This is the brute-force solution to the problem of differentiating characters, applicable in either statistics- or experience-based games (though fitting more easily into the former).

Character classes tell players, "Here's a problem. Your character can't solve this problem in ways X, Y, and Z because those methods aren't appropriate to your class. Find another solution that takes advantage of your character class's unique and clearly defined capabilities." Character classes are not a bad way to ensure experiential differences, but they're a little inelegant. I view them as a form of "remedial role-playing"; character classes have been a crutch for novice role-players since the 1970s. If reaching the mass market is your goal, character classes could be appropriate. If you're going for the Dungeons & Dragons audience, character classes are a necessity. I've noticed that the use of character classes is waning, but if the upcoming title Baldur's Gate proves as successful as most industry watchers expect, it could revitalize the notion of class distinctions in role-playing.

Varied Interaction

One of the defining characteristics of role-playing is the player's ability to impact the outcome of the story through his or her actions during the game. One of the most powerful and effective ways to give players this power is to offer and reward a variety of interaction styles.

When you play an Ultima, or even a hack-and-slash game such as Diablo, you meet computer-controlled people and creatures - NPCs - who -don't necessarily want to kill you. In a console game such as Suikoden, you build a base of operations and forge alliances with the NPCs you choose. Together, these activities have a dramatic impact. In almost any RPG you can name, players can kill, talk to, buy or sell from, maybe even learn from NPCs. The interaction can come in the form of one-sided info-dumps or in complex, branching tree conversations. Whatever form it takes, some nonviolent interaction is a necessary, defining characteristic of RPGs.

Furthermore, you can interact with the game environment in ways other than shooting weapons and opening doors. This may be as limited as picking up objects and manipulating switches and levers or it may be as limitless as, well, interactions in the real world. (I can dream, can't I?)

Finally, in the best RPGs, obstacles aren't limited to monsters or arbitrary puzzles. The solution to a particular game situation isn't as predictable as in more focused game categories and, often, more than one solution exists for each problem that you confront.

Combat

The effectiveness of your combat system depends on your understanding of the concept of "economy." Giving the player a big gun with unlimited ammo and then throwing a million hideous monsters at him or her -isn't going to buy you much sweat and adrenaline. However, try giving the player an ordinary gun with three shots and then send two villains at him or her and see what happens.

Make sure each weapon and each enemy is radically different than all others. There's no point offering a 1911 Colt, a Glock, and a Browning Hi-Power unless there's some major game play difference between them. Hey, they're all automatic pistols and serve about the same purpose in the real world. In your game, players -won't know that the stocks feel different, the weight is different, and so on. And tiny variations in accuracy and kickback and the like probably won't be noticeable even if you bother trying to simulate them. Don't bother. Just put one of them in your game and be done with it. If your game fiction can support it, make certain weapons particularly useful against certain enemies. Weapon differences must be obvious and instantly apparent.

Wherever possible, differentiate your enemies as much as (or more than) you do your player characters. If, again, your game fiction supports it, give each enemy an attack that has a specific effect (or effects) on the player's ability to move, see, hear, or otherwise interact with the game.

Do these, and you have a winning combat recipe. Because combat is easy to simulate on computers, and very little else is, I suspect (for better or worse) combat will remain a large part of RPGs, and much of our design effort will continue to go into crafting new combat systems.

Conversing With NPCs

No one has yet devised and/or implemented an artful, compelling, interesting, or believable conversation system in a computer RPG. That includes everything I've done and everything you've done. No one has come up with a system that -doesn't draw you out of the game world and remind you that you're just manipulating pixels on a screen. In the absence of anything better, let's look at some of the approaches that we've tried in the past.

First, there are branching-tree/keyword systems. If you've played just about any computer RPG of the last 15 years, you're familiar with these. Any Ultima game and, more recently, Fallout will introduce you to the concept, if you're unfamiliar with it (Figure 5). In this system, players read or listen to a bit of dialogue "spoken" by an NPC and are then offered a number of response options (or are given the opportunity to type in whatever they want). Picking one of these options or typing in a likely keyword sends the NPC into another speech. Making a selection typically prevents the player from getting the information he or she would have gotten by picking another of the available response options. Eventually, the NPC runs out of things to say along a particular branch and the conversation ends, leaving the player either to start the whole conversation over and make different response option choices in an attempt to elicit additional information from the NPC, or to go talk to someone else.

The problem is that clicking through a bunch of conversation options doesn't feel much like a conversation - an interrogation, perhaps, but not a conversation. Additionally, keywords and branching trees turn the conversations themselves into puzzles. Can you guess which branch the designer wanted you to go down? The opportunity and, more often, the necessity of talking to each NPC multiple times to be sure you ferreted out the critical nugget of information or set the one necessary conversation flag is a pain and drains conversations of their emotional impact.

Another way to handle NPC interaction is through linear conversations. This is sometimes called the "NPC as signpost" approach to conversation. It's most commonly used in console RPGs, where input options are limited and storage space for branching conversations is at a premium. Basically, this method boils down to walking up to an NPC and having them tell you something and that's it. No interaction. Talk to them again and, unless the game state has been advanced somehow and/or the designer is particularly sharp, the NPC will simply repeat what he or she said the last time.

Linear conversations typically point you to your next goal, but they can do much more. In the best examples, NPCs can tell you about themselves and their lives. They can describe in convincing terms how you know them and how they feel about you. It's possible to evoke real emotions in a linear conversation, and about all the writer has to worry about is the role the speaker plays in the story.

A third communication solution is the use of binary decision points. This is a compromise between branching and linear conversation approaches. Most often seen in console games such as Suikoden (Figure 6) and just about anything developed by Square, binary decision point conversations are linear except where a yes/no decision (and associated branch) will reveal something about the character, the player, or the NPC speaking. I think this is a most promising approach.

A fourth communication method is reaction-based conversations A few designers over the years have tried a system in which NPCs speak and the player gets to pick the tone of his or her response but not the specific content (wording) of the response. This -doesn't seem to offer much advantage over other, more popular systems, but it is an option.

The last communication option is simply denial. Back when Doug Church and I first started talking about System Shock, we were dissatisfied with the conversation approach taken in Underworld, traditional and conventional though it may have been. And though it pained us to admit it, even to ourselves, we had no idea how to do any better. So the team designed around the unsolvable problem - we killed everyone off. The inhabitants of Citadel station would exist, for the player, only through e-mail and video logs. It was an elegant solution to an intractable problem: if we can't make you believe you're talking to a real human being, we just won't have any in our game world. (In retrospect, I think we may have gone a little overboard - it was the right decision for that game at that time, but we failed to take into account the power of consistency and convention.)

Perhaps the best thing that can be said about conversation in computer gaming is that players have grown accustomed to inelegant, unrealistic, basically unbelievable systems and cardboard cut-out NPCs. Until someone comes up with something better, you can always fall back on convention, a fact that Doug and the System Shock team didn't consider very seriously. Players "get" branching-tree/keyword systems - they're so familiar with them that they -don't even think about them much anymore. And that's about the best that you can hope for - that conversation --won't drag players out of your carefully crafted alternate world too badly. I await the day when voice recognition, natural language processors, basic knowledge databases, and speech synthesis become realistic options.

Exploration Vs. Action

Having dealt with combat and conversation, there's only one defining characteristic of role-playing left to discuss: exploration. In the typical RPG, you can fully explore a huge, contiguous world in real time. Every Ultima, both Underworlds, Daggerfall, and countless other games have taken this approach. Uncovering all of a world's secrets is fun.

Clearly, the exploration model works in RPGs, but it's both a blessing and a curse. It's great to be able to say it takes a player hours to traverse your game's world (and that's if you don't stop to interact with anything), but that begs the question of whether it's any fun to walk around for ten hours. Additionally, game players are increasingly pressed for time or anxious to finish one game so they can move on to the next. As a result, many seem to want smaller worlds, shorter play times, and more frequent pats on the back. I know there are still hardcore game players out there who always want more, but you must decide whether there are enough of these consumers to support your development budget.

Recent trends favor smaller, deeply simulated worlds over large, contiguous spaces an inch deep and miles wide, at least in terms of depth of simulation. Further, many RPGs these days - my own Deus Ex, Thief: The Dark Project, Diablo, and others - are adopting a mission orientation. They break the world up into more manageable sections to minimize walking around and maximize fun. Mission structure also goes hand-in-hand with linearity and, together, they allow us to tell the best stories possible.

Variables

Now that we've discussed the defining characteristics - the rules of role-playing, if you will - what of the variables? There are certainly characteristics of role-playing that are not universal, that you can adopt or ignore, as you wish. Here are some of them.

CAMERA POSITION. I'm the last person qualified to address technology, but camera positioning is an important enough issue to cover briefly. There are several ways to approach camera positioning and player point of view in role-playing. The most common perspectives are: first-person, three-quarters-overhead third-person (as in, an isometric viewpoint), and top-down third-person. For the purposes of this discussion, and for the sake of brevity, I'll treat the third-person perspectives as the same.

In a genre that, at some level, boils down to providing the player with the Ultimate "I did this" experience, what could be more compelling than entering a new world and seeing it through your own eyes? The Underworld and Might & Magic series (Figure 8), not to mention many others, offer fine first-person role-playing experiences. If what you're after is simulating a world, a first-person view goes a long way toward making the player feel as though he or she is "there." You're reducing the distance between player and character to almost nothing.

One drawback to the first-person perspective is that it puts players at a tactical disadvantage by limiting their awareness of what's going on behind and to the sides of their characters. And if you're committed to turn-based, tactically challenging combat, convoluted conversations that take place on a separate conversation screen, and the video equivalent of paper game character sheets crawling with attributes, skills, and numbers, first-person could be the worst choice for your game. If you're trying to recreate the paper gaming experience or the experience you get when you read a great novel or watch a film - the fiction experience - a third-person view may be just what you need. To capture that "fiction feel," it's good to let players guide their characters rather than be their characters. You want a bit of distance between player and character.

A third-person perspective is also ideal for tactical decision-making, particularly when combined with a turn-based combat system, a character sheet, and a separate conversation screen. Each of these elements contributes to the player's ability to make informed decisions about how to develop his or her character. The trade-off is that it's tough to care much about NPCs that are obviously nothing more than bunches of pixels an inch high when you're looking at them from a bird's perspective.

Lone Adventurer Or Party?

If you're working on an RPG, there's one question I guarantee you've been asked: "Is there gonna be multiplayer support?" Everybody seems to want to go exploring with a party. In a papergame, where players gather around a table and engage in collective acts of imagination and push lead miniatures representing their characters around on tabletops, the party idea works just fine. Certainly the case could be made that allowing a party of adventurers go through your story together, linked via modem, LAN or the internet, is a worthy pursuit despite the problems of communication and coordination among party members.

But what of the single-player RPG? Does the party make sense in that context? Many classic RPGs indicate that it does. The early Ultimas and recent Krondor games (among many others) allow a single player to take control of a party of adventurers. However, I think controlling parties in single-player, story-based roleplaying game is a bad idea, particularly when it's a real-time game. There are several reasons I feel this way.

First, if one of the primary goals of role playing is to allow players to create an alter ego, you should do everything possible to increase a player's identification with his character. When controlling a party, player identification with a single character is history - as is roleplaying, in my opinion. At that point, you're playing a boardgame. Second, AI limitations mean you're inevitably going to be slowed down by teammates who can't think quickly on their feet and, even slowly, can't respond the way real people would.

If you subscribe to the ideal that players should describe their adventures by saying "I did" something rather than "Lara Croft did" something or "The Avatar" did something, controlling a party is a problem. In recent years, more and more designers seem to be coming around to this mode of thinking. (Recent Ultimas, Underworld, Diablo, and Daggerfall have all adopted the solo model, often to the chagrin of fans.) Solo play is simpler to implement, speeds up game play and fosters a direct connection between player and character that seems critical to RPG success.

If you choose to include full party control in your single-player game, recognize that you'll reduce player involvement and turn off people who value tactical thinking in games, and use it in conjunction with turn-based combat and a third-person perspective. A real-time, first-person party-based game is just asking for trouble.

Computer role-playing games are no longer just an infant medium learning to crawl. They've been around for a while,and now we, their designers, must start figuring out what we are, and what we can do to make the most of them. {Edit OK?} I've got no beef with folks who want to continue recreating the past. Just don't count me among the people who think that's good enough. Our goal should be to create games that are simple to learn and play, accessible to the broadest possible audience and yet with enough depth that hardcore gamers will flock to us.

The RPG Commandments